Skip to content

MULD2 Soil Map Parsers

What is this?

The MULD2 project (Mullastiku 2. ülevaatuskaart — Estonian Soil Map, 2nd revision) involves updating and quality-reviewing the Estonian national soil map. The original analogue map encodes soil properties as compact alphanumeric strings in three fields:

Field Estonian name What it describes
Siffer Šiffer Soil-type classification code(s) for the mapping polygon
Loimis Loimis Soil texture (particle-size composition) across up to 4 depth layers
Huumus Huumus Organic surface horizon — type, decomposition state, and thickness

These fields were authored over decades during analogue map digitisation. They carry a wide variety of encoding conventions, delimiter styles, and data-entry artefacts that must be resolved before the data can be used in quantitative analyses or database queries.

This library provides automated parsers that convert each raw field into a set of clean, typed database columns.

The three parsers at a glance

Parser Input example Key outputs
Siffer Ko;D;LPe siffer_1..4, n_siffers, parse_ok
Loimis l40-70/ls₂30/+ls₂ EST_TXT1..4, CLAY/SILT/SAND/ROCK 1..4, Z1..4
Huumus th15/h5 t₂20 h_type_1..4, h_depth_min/max_1..4, parse_ok

How the pipeline works

Each parser follows a two-phase approach:

  1. Repair — a lookup-driven pre-processing step corrects known digitisation artefacts (typos, wrong delimiters, legacy notations) before any grammar is applied.

  2. Parse — a formal grammar (Arpeggio PEG parser for Siffer and Loimis; a regex classifier for Huumus) validates the repaired string and extracts structured fields.

Results are written back to the PostgreSQL output_review table and exposed through this web application.

Live parser endpoints

You can test any code string directly in the browser:

Endpoint Route Description
Swagger UI /api/docs Interactive API explorer
Siffer parser /parse/siffer POST {"code": "Ko;LP"}
Loimis parser /parse/loimis POST {"code": "l40/ls₂"}
Huumus parser /parse/huumus POST {"code": "th15/h5"}

Or use the Muldade dekodeerija panel on the map page.

Where to go next